Traffic flow Theory

Traffic flow Theory
(Traffic Stream Characteristics)
Dr. Gang-Len Chang
Professor and Director of
Traffic Safety and Operations Lab.
University of Maryland-College Park
Traffic Flow Variables of Interest
Density (vehicles per unit distance)
S1
Spacing, or space headway between vehicles (distance per vehicle)
Time headway between vehicles (seconds)
Rates of flow (vehicles per unit time)
Dmi
Speeds (distance per unit time)
S2
S3
Travel time over a known length of road
S4
Occupancy (% of time a point on the road is occupied by vehicles)
Concentration (measured by density or occupancy)
1. Measurement capabilities for obtaining traffic data have been changed.
2. Measurement procedures estimated from variables will be discussed.
3. Uninterrupted flow facility: No external factors cause
periodic interruption to the traffic stream
Uninterrupted
Interrupted
Four Methods of obtaining traffic data (modified from Drew 1968, Figure 12.9)
Distance(m)
Along a Length of Road
Moving Observer Method
2500
2000
Measurement at a Point
1500
1000
Short Section
500
Freeway Lanes
0
30
60
90
120
Time(sec)
1
Measurement at a Point
Tube
Inductive Loop
Spot Speed
AGD
Systems
Direct measurements at a point: flow rate, headways, and speeds
Density, which is defined as vehicles per unit length, does not make sense
for using a point measurement.
A second observation location is necessary -> measurements over a short section
2
Over a Short Section
Paired Inductive Loop
Cameras
Paired inductive loops spaced five to six meters apart.
Provide direct measurement of volume, time headways, and speed.
Provide occupancy data by continuous reading, but does not permit
direct measurement of density.
3
Along a Length of Road
Several frames provide data for the estimation of speed and density.
Section 1
Section 2
Section 3
Section 4
Distance
It provides true journey times over a lengthy section of road, but that would require
better computer vision algorithms (few practical implementations)
Makagami et al. (1971), Persaud and Hurdle (1988b) constructed cumulative
arrival curves at several locations
average flow rate & average density within
a section, and consequently the average speeds through it.
4
Moving Observer Method
1 Floating car behaves as an average vehicle within the traffic stream
Method is effective for obtaining qualitative information about freeway operations
without the need for elaborate equipment or procedures
Method cannot give precise average speed data
2 Wardrop and Charlesworth (1954): obtain speed & volume measurements
Rural expressway data collection without automatic systems
Urban arterials with identifying progression speeds for coordinated signals
3 Wright (1973): Revisited the theory behind this method
Driver should fix the journey time in advance, and keep to it
Turning traffic (exiting or entering) can upset the calculations done using this method
Click to edit Master title style
1
(number of vehicles counted)
(flow rate)
(elapsed time)
(average headway)
Total elapsed time: sum of the headways for each vehicle
1985 HCM suggests using at least 15-minute intervals
(detail provided by 5-minute or 1-minute data is valuable)
Effect of different measurement intervals on the
nature of resulting data (Rothrock and Keefer,1957)
Different types of measured traffic flow or volumes
2
Time Mean Speed
Space Mean Speed
Arithmetic mean of the speeds of
vehicles passing a point on a highway
during an interval of time
Average speed of vehicles
measured at an instant of time
over a specified stretch of road.
A
speed (ft/sec or mi/hr)
1 n
ut = ∑ ui
n i =1
L=20km
A->B
B->A
80kph
40kph
L=20km
# of vehicles
B
(80) + (40)
TMS =
= 60kph
2
length of the
highway segment (km)
L
us =
n
ti
∑
i =1 n
number of
travel times
observed
SMS =
=
nL
n
∑t
i =1
i
travel time of
the ith vehicle
(hr)
20 + 20
= 53.3kph
3/ 4
2
More Flow-Concentration Work
Two mean speeds differ by the ratio of the variance to the SMS
ut = u s +
TMS
SMS
σ
2
us
variance to the SMS
s
( Wardrop ,1952 )
Breakdown from uncongested to stop & go conditions
: considerable difference
Gerlough and Huber (1975):
Relatively uniform flow and speeds, SMS & TMS are likely to be
equivalent for practical purposes.
Click to edit Master title style
2
Those systems that do not measure speeds(single-loop detector stations)
sometimes calculate speeds from flow and occupancy data (Athol, 1965)
Fraction of time
that vehicles are
over the detector
Vehicle
length
Occupancy =
specific time interval
If vehicle
lengths vary
and
Speeds
are constant
Detector
length
Vehicle
speed
=
=
=
=
=
=
3
Broader term encompassing density (measure of concentration over space) and
occupancy (measure concentration over time of the same vehicle stream)
Fractional shares
of total density
: Wardrop (1952), Gerlough and Huber (1975)
Average speed
Traffic stream : a number of sub-streams(constant spacing & speed)
Point measurements (single-loop) method strictly correct only under restricted conditions
Sub-streams with constant speed (Wardrop) is approximately true for uncongested traffic
Stochastic Real traffic flows: not be consistent with the measured data (Newell 1982)
3
Making more use of occupancy
“Density”
“Occupancy”
Vehicles per length of road
Directly affected by both of these
variables
Ignores effects of vehicle
more reliable indicator of the amount
length and traffic composition.
of a road being used by vehicles
Hurdle (1994) : the concept of acceleration in Newtonian physics
- There will be continued use of occupancy and density
1
(Jam density)

u

q = kj u −

uf

2
1
Speeds remain nearly constant even at quite high flow rates




Reduction in flow rates within the bottleneck at the time that
Uncongested
the queue forms upstream
2
Greenshields 1935
Queue Discharge
Congested
Hall, Banks (1991) & Hurdle (1988)
Uf : Free Flow Speed
4
HCM 1994 / HCM 2000
Hall and Montgomery (1993)
The breakpoints at which
Speeds remain flat as
speeds started to decrease
flow increases, and small
from free-flow, and
Decreases at capacity
113km/h
105km/h
97km/h
89km/h
the speeds at capacity
3
2
4
Speed (mph)
3 Hall and Brilon (1994)
1
Important for freeway management strategies
Will be of fundamental importance for ITS
implementation of alternate routing
Flow (veh/hr)
2
1
2
Greenshields 1935
Free Flow Speed

k
u = u f 1 −
 k
j

3
Huber 1957
Speed




k
u = c ln
k
 j
S = 38.05 - 0.2416D




Linear fit
Density (veh/km)
Speed (mph)
Jam Density
Speed (mph)
Speed (mph)
Jam Density
Density (veh/km)
Greenberg 1959
Logarithmic fit
Density (veh/km)
2
Edie's Hypothesis for the Speed-Density Function, Drake et al. (1967)
1
2
3
Empirical flow concentration data frequently have discontinuities in the vicinity
of what would be maximum flow
Need for Direct Flow-Concentration Relationship
Flow-concentration models can be derived from assumptions about the shape of the
speed-concentration curve
Duncan's (1979) : minor changes in the speed-density function led to major changes
in the speed flow function
3
Difficulty
With
Resulting
Models
(Ceder & May 1976,
Easa & May 1980)
The models often do not fit the data well at capacity
Little consistency in parameters from one location to another
Different days required quite different parameters
Use of volume & occupancy together to identify the onset of congestion
More
FlowConcentration
Work
Transitions between uncongested and congested operations at volume
slower than capacity
Use of time-traced plots to better understand the operations.
4
Three-Dimensional Relationship
(Gilchrist and Hall 1989)
+80km/hr
Catastrophe Theory
(Persaud and Hall 1989)
Further studied by Acha-Daza and Hall (1993), Pushkar et al. (1994)
70~80km/hr
A
B
C
60~70km/hr
D
50~60km/hr
E
~50km/hr
Five different colors represent different
Control variables(flow & occupancy) exhibit smooth continuous change
speed ranges.(Area A ~ E)
State variable (speed) undergo a sudden 'catastrophic' jump in its value
Conventional theory is insufficient to explain
the data
Further studied by Acha-Daza and Hall (1993), Pushkar et al. (1994)
It illustrates graphically that
1. Freeway operations do not have to stay on the curve;
jumps are possible from one branch to the other
2. Different locations will yield different types of data
Understanding
Traffic
Stream
behavior
Efforts
To
Implement
ITS
The relation between density and occupancy is weak.
Transforming variables, fitting equations, and then transforming
the equations back to the original variables can lead to biased
results
Provide challenges for applying this improvement in
understanding, how traffic operates, especially on freeways.
ITS will likely provide the opportunity for acquiring more
and better data to further advance understanding of these issues.